triage-deployment

Triage CI Failures

Diagnose failing GitHub Actions workflows in this codebase, map errors to their root causes, read the relevant configuration, apply targeted fixes, and verify locally.

When to Use

• A CI check is red on a pull request or push
• The user pastes GitHub Actions log output and asks for a fix
• A gh run or workflow URL is provided for inspection
• Asked to "fix CI", "debug the failing check", or "triage CI errors"

Step 1: Obtain the Logs

Option A — Specific logs provided in context

If the user has pasted log output or given a specific run URL in $ARGUMENTS, fetch only that run and skip to Step 2:

# e.g. https://github.com/JimmyPaolini/codebase/actions/runs/12345678
gh run view 12345678 --log-failed

Option B — No logs provided: fetch ALL failing runs for the current PR

When no logs are provided, list all failing checks on the current PR and retrieve logs for each. Work through each failure in sequence — do not stop after the first fix.

# List all failing checks on the current PR (name + run URL)
gh pr checks --json name,state,link \
  --jq '.[] | select(.state == "FAILURE") | "\(.name) \(.link)"'

Parse the run ID from each link (last path segment of the URL) and fetch its logs:

gh run view <run-id> --log-failed

For each failing run, fetch its logs:

gh run view <run-id> --log-failed

Process all failing runs before moving to Step 4. Each failure may require a separate fix — apply them all before verifying.

Step 2: Identify the Workflow and Failing Job

Match the log header against the known workflows:

Workflow name	Job name	Trigger
🧑‍💻 Lint Codebase	lint-codebase	push / PR / manual
🧑‍🔬 Test Coverage	test-coverage	push / PR / manual
🧑‍⚖️ Validate Conventions	validate-conventions	PR (opened/sync/edited) / push to main
👮 Audit Issues	audit-issues	issue opened/edited/labeled/unlabeled
🕵️ Scan Security	scan-security	push / PR / weekly schedule